Why does the holiday take effect?
Timing of cesarean section is a daily regular waving pattern that had been documented by Joanne Spetz[1]. It might also be influenced by financial incentive and physicians' convenience.
However, the CS’s technique for obstetricians is simple. But they are still cautious in each decision making of CS.Financial incentive and time convenience might be a part of determinants in timing. There was a weekly regular waving pattern that total number of births jumps up in weekdays and falls down in weekends. And Cesarean section (CS) does in a similar way. Obstetricians could not practice alone. While practicing either CS or vaginal delivery (VD), they should be supported with their hospital facilities and other specialties. Therefore, the timing and schedule of CS could be affected with the facilities constraint and other specialties’assistance.
What a physician practices in a hospital should be constrained with capacity of his/her hospital[2]. In an open system,obstetricians would contract with hospitals to have the admitting privileges. Their services might be constrained by the hospital resource allocation that is plentiful in weekdays but confined in weekend. On the other hand, obstetricians are employed by their hospitals in a closed system. All obstetricians were obligated to follow some hospitals’ policy; they could not operate their CS conveniently as they wished.
In a hospital, obstetrics department operates all year round. Usually,the support in office hours, weekdays and working days is better than that in off hours, weekends and holidays. In other words, there is a regular waving pattern that the capacity of a hospital is sufficient in office hours, weekdays and working days but the rest is scant. It means that the obstetric department provides only a basic resource level to maintain its operation during off hours, weekends and holidays.
Only urgent medical servicescan not be scheduled. There are many services such as the labor induction and a part of CSs that are flexible in timing. However, physicians would take some risk in consideration about working with limited resource. Obstetricians would bring some non-urgent and CSs scheduled easily forth or postpone them. This way would make the limited resource that could be preserved for urgent or unexpected deliveriesin off hours.
Reviewing the literature and the author's experience, some indication such as previous CS, breech and some kind of medical diseases could be scheduled. The cases of fetal distress and pelvic dystocia, occurring in trying labor, are not expected. On the other hand, labor induction are seldom performed at Friday or weekend to prevent urgent CSs. So, pregnant women are usually advised for labor inductions at weekdays except Friday. After doing as above, the numbers of both total births and CSs would be plottedin an M type waving pattern.
The scant resource during holidays (5 days to 10 days) is also noted in hospitals. Physicians would advise their patients to shift the schedule earlier or later. The waving pattern of CS numbers would be similar to that in weekend.
Hypotheses
The services of both CS and VD are constrained by supports of hospital facilities and assistance of faculties. In fact, the weekend supports are usually less than the weekdays. Then the hypotheses A1 and A2 are proposed.
Hypothesis A1: The numbers of birth in weekdaysare higher those in weekends.
Hypothesis A2: The numbers of CS in weekdays are higher those in weekends.
The CSs need more human resource and facilities such as operation rooms, equipments, anesthesiologists and other faculties. Comparing with a VD, some CSs could be scheduled easier. While availabilities and facilities in hospital are constrained in weekends or holidays, obstetrician would bring some CSs forward or postpone them.
That some CSs seem to be scheduled easily then the hypotheses B1 and B2 could be implicated.
Hypothesis B1: Among weekdays and weekends, the viability of CS is more prominent than that of vaginal delivery.
Hypothesis B2: Among office days and holidays, the viability of CS is more prominent than that of vaginal delivery.
Some CSs are indicated such as breech, previous CS, placenta previa, and medical complication could be advised to have their operations. Physicians could schedule them regularly. These kinds of CSs in this research are defined as “Schedulable CS”. Other indications like fetal distress or dystocia (Prolonged Labor) could be categorized as “Un-Schedulable CS”.
Frankly, “Un-Schedulable CS” can not be scheduled. These CS cases would distribute in each day randomly. “Schedulable CS” would be frequently arranged that the case numbers would be concentrated in weekdays. Like in weekend, the case number of “Schedulable CS” would be less among the long holidays.
Some indications for schedulable (regular) CS are frequently scheduled in previous. They would be arranged in weekdays. But other indications for un-schedulable (urgent) CS could not be scheduled previously[3]. They should spread around weekdays and weekends. The viability could be different for two groups. Thus hypotheses C1 and C2 are shown.
Hypothesis C1: Between weekdays and weekend, the viability of schedulable CS is more prominent than that of un-schedulable CS
Hypothesis C2: Between office days and holidays, the viability of schedulable CS is more prominent than that of un-schedulable CS.
Private hospitals have stronger incentives to keep better facilities in weekends or holidays. But public one does not. In private hospitals, the scheduled CSs could be arranged in weekends and holidays. Then hypotheses D1 and D2 would be presumed.
Hypothesis D1: Among weekdays and weekend, the viability of schedulable CS in a public hospital is more prominent than that in private one.
Hypothesis D2: Among office days and holidays, the viability of schedulable CS in a public hospital is more prominent than that in private one.
Data and Methodology
Based on the NHI service records since 2003 to 2005, all the cases of birth, CS and VD, had been abstracted and analyzed.
According to ICD-9 records, two indications such as dystocia (Disproportion653, Obstructed labor 660, Abnormality of forces of labor 661 except 661.3, Long labor 662, Malpresentation 652 except 652.1 and 652.5, Failed induction of labor 659.0 or 659.1) , fetal distress (656.3) and cord prolapse(663.0) are categorized as un-schedulable CS. Other CS indications such as previous CS (654.2), breech (652.2) and some medical diseases are categorized as schedulable CS[3].
This study had defined a weekly waving index as following
Represent the case number of Mon, Tue, Wed, Thu, Fri, Sat and Sun respectively within a week.
There were two groups VD and CS. Each had its 52-week waving observations around a year. This study used the independent-sample T test to testify whether VD and CS had different waving or not (Hypothesis B1). Also, there were two groups (Schedulable CS & Un-schedulable CS) use this method to testify the different waving between them (hypothesis C1).
Then, two groups (“Schedulable CS” between public & private hospitals) had repeated this method to testify hypothesis D1.
There were different length of holidays in 2003, 2004 and 2005. Each length was 7, 7 and 9 days according to the government’s annual calendar. Then, the length of “Before Holidays” and “After Holidays” were defined as 7 days for each year. We calculated the following odds ratio
to testify that the viabilities of CS and VD were different (Hypothesis B2). Also, we repeat the same way to testify hypothesis C2 and D2.
Hypothesis C2
Hypothesis D2 (schedulable CS)
Hypothesis D2’ (un-schedulable CS)
Results
Tables (1,2,3), Figure 2
(A1 and A2)(B1 and B2)(C1 and C2)(D1 and D2)
The statistical results of hypotheses A1 and A2 are presented in table 1. The significant values t-statistic could show that the average numbers per day of total birth and CS in weekdays arehigher than those in weekends from 2003 to 2005. Regarding with these results, the numbers of total birth and CS would circle weekly.
The statistic result (Table 2, B1) showed that waving indexes CS and VD are significantly different among weekdays and weekends.The viability of CS is more prominent than that of vaginal delivery (Figure 2, B1). Among the office days and holidays, the odds ratio results (Table 3, B2) in three years are below 1. The occurrences rate of CS within holidays was less than that before or after holidays. Comparing with VD, the CS contributes more for the number’s viability (Figure 2, B2).
There was no significant difference between the weekly waving curves of schedulable and un-schedulable CS. They seem similar (Figure 2, C1). The statistical results also showed that waving indexes of them were not always different in three years (Table 2, C1). The curve of the schedulable and un-schedulable case numbers waved simultaneously (Figure 2, C2). The odds ratio results are around 1 in three years. It could not support hypothesis C2 that they are different.
The weekly curves of schedulable CS in public and private hospitals moved inconsistently (Figure 2, D1). The result of waving indexes showed a significant difference in three years. The viability of public hospitals was more prominent than that of private hospitals (Table 2, D1). In contrast with D1, both public and private hospitals’ curves of schedule CS numbers showed a similar M type (Figure 2, D2). There are no significant statistic results to support hypothesis D2.
Discussion
Waving of Birth (CS and VD)
Regarding with the figure 2 (B2, C2 & D2), they showed an M type curve. Why were the CSs operated within holidays less than those before or after holidays? The limited resource within holidays makesobstetricians arrange their schedules away from holidays. Either bringing them forward or postponing them, it makes an M type waving curve.
An obstetrician would not arrange a schedule of CS and induction in weekends and holidays to avoid their own malpractice risk as possible as he/she can. Comparing with a spontaneous delivery, CS and induction need teamwork and the more delivery room facilities. These resources are limited within holidays and weekends. While lacking, the risk of malpractice might be higher. Obstetricians would not take this risk and they would rationally schedule some CSs and inductions in a time period that resources are not limited. The distribution of CS occurrences might seem as an M type curve around holidays. Owing to limited resources in weekend, there was a weekly waving of CS.
Inthis article’s reasoning, an obstetrician could not arrange un-schedulable CS as his/her will. The occurrence rate of un-schedulable CS would not correspond with the resources allocated by hospitals. But our results C1 and C2 could not be supported. Not only the schedulable CS but also the un-schedulable waved simultaneously. The results are that time convenience and resource allocation could even influence on time arrangement of un-schedulable CSs.
Schedulable and Un-Schedulable
Isomorphism (C1, C2)
Obstetricians would advise pregnant women for a proper way of delivery. According to patients' indications, they can be divided into two groups. One group is high risk for VD, these women such as previous CS, breech and some medical problemwould be advised for schedulable CS. The Other group is low risk women who could try spontaneous labor or induction with uterine contraction agentto accelerate their processing.
During their labor, these low risk pregnant women might suffer from two major complications such as fetal distress or dystocia. Then they could be advised to operate emergently.
Obstetricians would arrange the schedulable C/S at weekdays. However, the some postdatepregnancies or macrosomia areadvisedfor induction service at office days (W1-W4). Friday seems not to be a good day for induction because they could meet the fetal distress or dystocia within weekends. Then, the emergentlyCS from these inductions, which are categorized as the un-schedulable, would be away from weekends and holidays. As the results of C1 and C2, the schedulable and un-schedulable CSs have a similar waving.
Resource Sufficiency
(D1, D2)
Influence of resource allocation
Private hospitals have a stronger incentive to maintain better services. They would allocate more resources within weekends. It means that obstetrician could arrange their CSs within weekends.But the public hospitals have a short supply within weekends. The waving of schedulable CS between public and private hospitals was different (Result D1).
But in long holidays of Chinese New Years, that is same as American Thanksgiving Day and Christmas. Within these holidays, all families would get together to celebrate them. In Both public and private hospitals, obstetrician and staffs would be off. They do not have the incentive to maintain their service within the holidays. The facilities and service teams would be limited to a basic running in both types of hospitals. There would be no difference between them (Result D2).
Those un-schedulable CSs had similar results as the schedulable ( Table 2, D1’ & Table 3, D2’). They are illustrated by Figure 2, D1’ and D2’.The hospitals’ resource sufficiency in weekends and within holidays would even influence on the un-schedulable CSs. It reinforces that resource constraints in a hospital could be a determinant factor for an obstetrician to schedule their un-schedulable CSs.
.
Not only physician time selection but also resource constraints in hospitals would influence on CS schedule.
Table 1. Statistical Results of Hypotheses A1 and A2
A1 / A2Mean
(Std. Deviation) / T / Mean
(Std. Deviation) / T
2003 / Weekday
(N=261) / 618.57
(70.25) / 11.210*** / 210.76
(48.94) / 10.002***
Weekend
(N=104) / 544.33
(50.95) / 167.83
(31.02)
2004 / Weekday
(N=262) / 593.64
(73.68) / 8.325*** / 195.33
(45.19) / 7.591***
Weekend
(N=104) / 531.63
(60.13) / 161.22
(36.46)
2005 / Weekday
(N=260) / 562.04
(74.70) / 6.683*** / 187.44
(45.34) / 6.380***
Weekend
(N=105) / 503.32
(76.50) / 156.54
(40.40)
***p<0.001
Table 2. Statistical Results of Hypotheses B1, C1 and D1
N=52 / 2003 / 2004 / 2005Mean
(Std. Deviation) / T / Mean
(Std. Deviation) / T / Mean
(Std. Deviation) / T
B1 / Waving of CS / 1.87
(0.39) / 11.610*** / 1.81
(0.39) / 10.810*** / 1.76
(0.40) / 9.539***
Waving of VD / 1.22
(0.09) / 1.22
(0.06) / 1.22
(0.08)
C1 / Waving of schedulable CS / 2.01
(0.49) / 2.145* / 1.97
(0.49) / 2.691** / 1.89
(0.52) / 1.370
Waving of un-schedulable CS / 1.83
(0.39) / 1.73
(0.39) / 1.76
(0.40)
D1 / Waving of schedulable CS in public hospitals / 4.10
(3.26) / 4.626** / 3.70
(1.94) / 6.309** / 4.51
(5.01) / 3.774**
Waving of schedulable CS in private hospitals / 1.98
(0.53) / 1.94
(0.52) / 1.87
(0.53)
D1’ / Waving of un-schedulable CS in public hospitals / 3.35
(2.25) / 4.765** / 2.93
(1.25) / 6.536** / 3.66
(2.39) / 5.688**
Waving of un-schedulable CS in private hospitals / 1.83
(0.44) / 1.74
(0.41) / 1.75
(0.40)
***p<0.001
*p<0.05, **p<0.01
Table 3. Statistical Results of Hypotheses B2, C2 and D2
Within Holidays vs.Before Holidays / Within Holidays vs.
After Holidays
2003 / 2004 / 2005 / 2003 / 2004 / 2005
B2 / odds ratio / 0.714*** / 0.712*** / 0.799*** / 0.617*** / 0.670*** / 0.677***
95% C.I. / 0.648~
0.785 / 0.645~
0.786 / 0.727~
0.878 / 0.560~
0.679 / 0.607~
0.740 / 0.617~
0.743
C2 / odds ratio / 0.934 / 1.036 / 0.954 / 0.970 / 1.036 / 1.134
95% C.I. / 0.795~
1.098 / 0.877~
1.224 / 0.815~
1.118 / 0.827~
1.139 / 0.877~
1.223 / 0.937~
1.322
D2 / odds ratio / 10.870 / 1.006 / 0.847 / 0.969 / 1.044 / 1.083
95% C.I. / 0.623~
1.214 / 0.707~
1.432 / 0.607~
1.180 / 0.693~
1.353 / 0.734~
1.487 / 0.780~
1.506
D2’ / odds ratio / 1.02 / 1.174 / 1.050 / 0.906 / 1.186 / 0.926
95% C.I. / 0.697~
1.496 / 0.809~
1.703 / 0.716~
1.541 / 0.624~
1.317 / 0.818~
1.718 / 0.640~
1.340
***p<0.001
Figure 1
Figure 2
2003 / 2004 / 2005B1 / / /
B2 / / /
C1 / / /
C2 / / /
D1 / / /
D1’ / / /
D2 / / /
D2’ / / /
1.Spetz, J., Mark W. Smith and Sean F. Ennis, Physician Incentives and the Timing of Cesarean Sections: Evidence from California. Medical Care, 2001. 39(6): p. 536-550.
2.Harris, J.E., The Internal Organization of Hospitals: Some Economic Implications. The Bell Journal of Economics, 1977. 80(2): p. 467-482.
3.Olivia A. Henry, M., BCh, MPH, Kimbery D. Gregory, MD, MPH, Calvin J. Hobel, MD, and Lawrence D. Platt, MD, Using ICD-9 Codes to Identify Indications for Primary and Repeat Cesarean Section. American Journal of Public Health, 1885. 85(8): p. 1143-1146.